Gmchimica
et Cosmochrmica
Am Vol. 56, pp. 4099-4103
Copyright Q 1992 Pergamon
Prffs
Ltd.
Pnnted in U.S.A.
0016.7037/92/$5.00
+ 40
LETTER
Rhenium in seawater: Confirmation of generally conservative behavior
A. D. ANBAR, R. A. CREASER,*D. A. PAPANASTASSIOU,
and G. J. WASSERBURG
The Lunatic Asylum of the Charles Arms Laboratory, Division of Geological and Planetary Sciences,
California Institute of Technology, Pasadena, CA 9 1125, USA
(Received May 12, 1992; accepted in revisedform August 25, 1992 )
Abstract-A
depth profile ofthe con~ntration of Re was measured in the Pacific Ocean using a technique
we have developed for the clean chemical separation and the precise measurement of Re by isotope
dilution and negative thermal ionization mass spectrometry (ID-NTIMS). This technique permits Re
concentrations to be determined from 200 mL of seawater with a typical precision of +5%0. This is an
improvement of at least a factor of 100 over the techniques used in previously published determinations
of Re in seawater. We obtain a narrow range for Re from 7.20 & 0.03 to 7.38 f 0.03 ng/kg for depths
between 45 m and 4700 m. This demonstrates that Re is relatively well mixed throu~out
the water
column and confirms the theoretical prediction that the behavior of Re in the oceans is conservative.
When examined in detail, both salinity and the concentration of Re increase by approximately 1.5%
between 400 and 4700 m, a correlation consistent with conservative behavior. However, Re appears to,
be depleted relative to salinity by l.O-1.5% at 100 m, and enriched by approximately 4% at the surface.
These observations suggest a minor level of Re scavenging in near surface waters, and an input of Re to
the ocean surface. This work demonstrates the utility of ID-NTIMS for geochemical investigations of
certain trace elements that have not previously been amenable to detailed study.
in seawater indicate a range of approximately 3 to 11 ng j kg
(SCADDEN, 1969;OLAFSON and RILEY, 1972; MATTHEWS
and RILEY, 1970; KOIDE et al., 1987). These variations
probably reveal a lack of precision and sensitivity in the analytical methods used, so that the content of Re in seawater
is determined by these data to no better than a factor of three.
Our understanding of the geochemistry of Re and the PGE
is limited largely by their low abundances and, until recently,
the lack of analytical techniques which are sensitive, precise,
and accurate enough to obtain reliable abundance measurements, The application of NTIMS to these elements ( ZEININGER,1984, DELA.IORE,
1987; HEUM~,
1988; VUL~~NG
et al., 199 1; CREASERet al., 1991) introduced a highly sensitive new tool for their study. In order to resolve the problem
of the Re concentration of seawater and its distribution with
depth, we have applied ID-NTIMS to the study of Re in a
Pacific Ocean profile.
INTRODUCTION
THE GEOCHEMISTRY
OF the Platinum Group Elements ( PGE)
and their periodic table neighbors is poorly understood, despite increasing interest in recent years in the use of these
elements as isotopic and elemental tracers, and for geochronometry. This problem is particularly acute in the marine
environment, where the surface abundances of OS, Ir, Ru,
and Rh are poorly constrained, and their depth profiles are
unknown. Depth profiles have been measured for Pt, Pd, and
Re (e.g., Pt: GOLDBERGet al., 1986; JACINTOand VAN DEN
BERG, 1989; Pd: LEE, 1983; Re: MATTHEWSand RILEY, 1970;
OLA~SON and RILEY, 1972; KOIDE et al., 1987), but these
data have raised as many questions as they have answered.
For example, the studies of the marine distribution of Pt
have each yielded a different trend with depth. The cause of
these differences is undetermined.
Uncertainty has also surrounded the marine geochemistry
of Re. The concentmtion of Re in oxygenated seawater is
expected to be fairly uniform with depth, as the stable form
of Re in these waters should be the perrhenate anion,
ReO; (BRULAND, 1983; BROOKINS, 1986). Thus, the behavior of Re should be conservative in seawater. This expectation is consistent with the apparent enrichment of Re
in seawater relative to its periodic table neighbors ( KOIDEet
al., 1987), which suggests that Re has an extremely long residence time in the oceans (GOLDBERGet al., 1988). However,
the published measurements of the concentration of rhenium
SAMPLING AND ANALYSIS
Seawater samples were collected in January 1992, on the HOT33 cruise (R/V Wecoma) at Station Aloha, approximately 100 km
north of Oahu (22”N, 158’W). Nine samples were collected using
PVC Niskin bottles with Teflon-coated internal springs, lowered on
a steel cable. One sample (at 45 m) was collected using a Go-Flo
bottle on a Kevlar line. Pressure, tem~mture, conductivity, and dissolved oxygen were monitored continuously using a conductivitytemperature-depth probe (CTD) and an oxygen sensor (KARL and
WINN, 199 1) . The depth, salinity, and oxygen data reported in this
study were derived from the CTD and oxygen sensor data (Table 1;
D. Karl and R. Lukas, pers. commun.). After recovery, all samples
* Present add~e~~~
~~rnent
of Geology, University of Alberta,
l-26 Earth Sciences Building Edmonton, Alberta, Canada T6G 2E3.
4099
were filtered within 48 h through 0.2 gm n&cellulose filters using
a peristaltic pump, then acidified by adding 2 mL of 12 N HCl to
each liter of seawater, and stored in acid-washed high density polyethylene bottles for later analysis.
4100
A. D. Anbar et al.
TABLE 1.
Reconcentrations in the Pacific Ocean
Sample
Re
BU!l~fE@!&(El&02e
ReN
7.70
7.70
7.66
6(a)
7.68
45(b’
7.38
51
7.35
100
402(a)
749(b’
1499
2196
2600(a)
4705(b)
7.34
7.20
7.24
7.30
7.31
7.32
7.34
7.36
7.34
7.35
7.33
7.39
7.43
7.44
7.40
7.40
7.41
7.44
7.41
Salinity Oxygen
mlumol/ka)
0.07
0.06
0.03
0.03
0.03
0.03
0.04
0.03
0.03
0.03
0.06
0.03
0.03
35.002
205.4
35.006
35.089
34.104
205.4
205.4
156.3
26.8
“3”4:35::
:3::
34.622
111.6
34.653
34.645
138.4
ReN is the Re concentration normalized to 35 % salinity.
The uncertainty of each measurement is derived from
uncertainties in the measured 249/251 ratio, tracer concentration, and all gravimetric determinations, (see text).
Replicate samples (6, 402 and 2600 m) were processed
separately, and indicate that the anal lcal variability was
within the uncertainties of the indivlI? ual measurements.
Samples 6 (a), (b) and 2600 (b) were spiked using the Dil2 tracer. All other samples were spiked with Dil-1.
Accurately weighed 200 mL fractions from all samples, except for
the 6 m sample, were spiked with a ‘s5Re-emiched tracer solution
approximately 3 months after collection. The 6-m sample was spiked
aboard ship, following filtration and acidification. In all spiked solutions, the measured ‘*‘Re/ “‘Re ratio was close to the optimal
value needed to minimize the error propagation factor. After the
addition of the tracer, the solutions were periodically shaken and
stirred over 24-48 h, which was deemed sufficient time for complete
equilibration of the Re isotopes, since Re is presumed to be present
as ReO; in both samples and tracer. The Re tracer was prepared by
dissolving ‘85Re-enriched metal (Oak Ridge National Laboratory)
in 16 N HNO, and diluting to I N with ultrapure water. Two dilutions
of this stock solution were prepared. Dilution 1 ( Dil- 1) was used to
spike the 6 m sample aboard ship, and a 2800 m fraction in the lab.
Dilution 2 (Dil-2) was used for the other samples spiked in the laboratory. The concentration of la5Re in both of these tracer solutions
was determined by calibration against a gravimetrically prepared
standard solution of isotopically normal, high purity Re metal
(Johnson Matthey). The isotopic composition of the tracer solutions
was determined by NTIMS in our laboratory (‘s’Re/‘*‘Re = 37.14
+ 0.04), and the isotopic composition of the standard solutions and
natural samples is assumed to be ‘ssRe/‘87Re = 0.5974 + 0.0004
(GRAMLICH et al., 1973). The isotopic composition of Re standard
solutions determined by NTIMS in our laboratory (0.5976 + 0.0003)
is identical to the isotopic abundance ratio determined by GRAMLICH
et al. ( 1973), using carefully controlled procedures.
Rhenium was separated from the seawater samples by anion exchange chromatography using two small Teflon columns (e.g.,
HUFFMANet al., 1956). Primary separation utilized 200 PL Bio-Rad
AGI-X8,100-200 mesh resin, onto which ReO; wasadsorbeddirectly
from the spiked, acidified seawater. The column was rinsed with 10
mL ultrapure H20, followed by 650 pL 4 N HN03 to remove MO
and W. These elements are both present as oxyanions in seawater at
higher concentrations than Re. Rhenium was then eluted with 2 mL
4 N HNOB. The resulting solution was evaporated to a small drop
at 70-80°C under N2, and passed through a second column with 510 pL of resin to purify Re further and to reduce the small amount
of organic residue remaining from the first column. This entire se.p
aration procedure had a yield of 90%, determined by isotope dilution
analysis of a known amount of Re that was passed through the chemistry. The chemical blanks were in the range 2-5 pg. For all the data
reported below, the quantity 3.5 pg was subtracted from the quantity
of sample Re that was determined to be present by isotope dilution
analysis. Since approximately 1.5 ng of Re was typically separated
from the Seawater aliquots, the resulting blank correction was less
than 3%0of the total sample Re present. This is smaller than the total
uncertainty of the analytical procedure (see the following text).
The chemically separated Re was loaded using polyethylene tubing
on a microsyringe, onto a high purity Pt filament on which Ba had
been loaded as BaS04 (used to enhance negative ion emission), and
analyzed in a Lunatic mass spectrometer configured for negative ions
( WASSERBURGet al., 1969; CREASERet al., 199 1) . When prepared
as a fine precipitate, BaSO., forms an even coat on the filament, on
which the sample may be easily loaded. The BaSO, was found to
provide a reproducible and efficient method of obtaining a higher
ionization efficiency for both samples and standards than was possible
using other Ba salts, such as Ba( (NO)& or Ba( OH), (CREASERet
al., 1991; HEUMANN, 1988). The ionization_efficiency for Re standards was 20%. The ionization efficiency was lower for separated
samples due to the presence of organic material in the final load, and
ranged from 1 to 10% for the samples analyzed in this study.
Re was analyzed as ReO; at masses 249 and 25 1, and the ion
intensity at 25 1 was corrected for the contribution from “‘Re I603l8Ousing ‘sO/‘60 = 0.002085 to obtain the ratio ‘s5Re/187Re. This ‘*O/
I60 ratio has been measured directly for normal Re, using the
‘*‘ReO; ion beams (CREASER
et al., 1991). The contribution from
‘87Re’702’60; at mass 25 1 is negligible. When running seawater
samples, ion currents corresponding to lo’-10’ counts per second
(cps; 0.16-1.6 X lo-” A) at masses 249 and 251 were routinely
sustained for 30-40 min on a Faraday collector, at filament temperatures of 840-89O”C, as determined by an optical pyrometer.
To determine the level of the ion current due to background contamination, BaS04 was loaded on a filament without any sample or
tracer. Ion currents corresponding to lo’-lo4 cps at mass 251 were
measured over the temperature range 840-890°C. This current increased with increasing filament temperature, but was fairly constant
over time when the temperature was not varied. The ratio ‘*‘Re/
‘*‘Re measured in this experiment was indistinguishable from that
of normal Re.
In order to establish the level of the filament blank under controlled
conditions, we loaded 36 pg of Re tracer on a Pt filament with BaSO4,
and obtained a stable signal at mass 249 equivalent to = lo6 cps, at
a filament temperature of 850°C. The initial *8sRe/‘87Re measured
was 33.4, corresponding to a blank equivalent to 0.17 pg of normal
Re from the filament. The run was continued at constant temperature
until approximately 3% of the initially loaded Re tracer had been
ionized and detected (determined by integration of the ion beam
intensities over time). If a constant, but initially unknown, blank
contribution to the measured ion beams is assumed, then the data
from this experiment can be used to determine the magnitude of the
blank contribution. This contribution corresponds to ~5 X lo3 cps
at mass 251.
The Re blank contributions determined from these two experiments
are in general agreement. It follows that for our current generation
of “clean” Pt filaments, the background current, using BaSOd as the
emitter, corresponds to 103-lo4 cps at mass 25 1. This low level of
Re blank has broad applicability in Re-0s geochemical studies.
During analyses of seawater samples, the observed drift in the
measured ratios led to a typical uncertainty in these ratios of less
than 21.5% amu-‘, consistent with mass dependent isotope fmctionation associated with sample depletion. No other drift was observed (e.g., as would be obtained from a variable filament blank),
even when beam intensities were increased by an order of magnitude
under different operating conditions. Replicate analyses of samples
at 6 m, 402 m, and 2800 m were in agreement to +3k (Table l),
consistent with this source of uncertainty. The close agreement of
the two 2800 m fractions, spiked with different tracer solutions, indicates that the calibrations of the Dil-1 and Dil-2 tracer solutions
are well within the analytical uncertainties.
An uncertainty propagation calculation was made for each sample,
including the uncertainties in all gravimetric measurements, as well
as the uncertainties in the concentration and isotopic composition
of the Re tracer solution, and the uncertainty in the isotopic composition of natural Re. An uncertainty of +3L was assigned to the
measured isotopic ratio of each spiked sample, due to the mass-fmctionation uncertainty discussed above. An uncertainty of rt50% was
assigned to the 3 pg chemical blank. The cumulative uncertainties
from these sources (at 95% confidence) are reported in Table 1, and
Marine geochemistry of rhenium
0
I
1000
E
2000
--
dilution-indu~ively
coupled plasma mass spectrometry
(ICP-MS), which yield marine profiles that are uniform with
depth to within 5~4%( COLODNER, 199 1).
The consistency of our data is a significant improvement
over the previous measurements of Re in seawater. Our best
estimate of the average Re concentration of seawater is 7.3 1
+ 0.11 ng/ kg, or 7.42 & 0.04 ng/ kg on a salinity-normalized
basis. These values were obtained by averaging all the data
below 400 m, where the salinity-normalized Re concentrations are most uniform. The uncertainties are f2a of the
data used to calculate the averages. The samples from depths
shallower than 400 m appear to deviate somewhat fi-om conservative behavior, but it is not clear whether these deviations
represent actual variations in the concentration of Re in the
water column (see the following text and Fig. 2). Thus, these
measurements are omitted from the best estimate. Table 2
compares our best estimate of the concentration of Re in
seawater with the corresponding estimates from other studies.
Our determination is within the wide range of the analyses
of SCADDEN( 1969), MATTHEWS and RILEY ( 1970), and
KOIDE et al. ( 1987), but is substantially higher than the value
from the Atlantic Ocean study of OLAF~SON and RILEY
( 1972). As suggested by KOIDE et al. ( 1987), we suspect that
the uncertainties in these earlier studies are due, in large part,
to variability in the yields of the chemical separation of Re
from seawater. A systematically low yield could also explain
the shift between our average value and that of OLAI?SSON
and RILEY ( 1972). Since, in our isotope dilution analyses,
the tracer was added prior to chemical separation, yield uncertainties were not a source of error, as long as the sample
and tracer isotopes equilibrated. The reproducibility of our
data provide support that this was the case.
Our average seawater value is significantly lower than the
best estimate by COLODNER( 1991) of 8.19 + 0.37 ng/kg,
t
5
4
0
3000
0
I
FIG. 1. The concentration of Re at various depths in the Pacific
Ocean. The solid circles are data from this study; other points are
from KOIDEet al. ( 1987). Our samples were taken at 22”N 158”W.
Other samples were taken at 33”N 139”W (open circles); 35”N
122OW (triangles), and 35’N 138”W (crosses).
were typically within +5k. This is consistent with mass fractionation
as the primary source of uncertainty for most samples.
RESULTS AND DISCUSSION
The results are shown in Table 1. The Re concen~tions
are presented both as direct dete~nation~
and on a salinitynormalized basis (normalized Re, Re = Rhm,l, X 35 /Sample salinity). A comparison of the depth profile determined
in this study with the Pacific Ocean profiles of KOIDE et al.
( 1987 ) is shown in Fig. 1. In contrast to the previous studies,
our data clearly demonstrate that the concentration of Re is
essentially constant with depth, and that the chemistry of Re
is generally conservative, as was predicted on thermodynamic
grounds ( BRULAND, 1983; BROOKINS, 1986). This result is
consistent
with less precise recent measurements
32.5
33.3
using isotope
34.0
34.7
4101
35.5
0
0
1000
1000
2000
2000
E
ZE
0
0 3000
a
f
g
3000
4000
4000
a
5000
7.00
7.20
7.40
7.60
Re Ma91
7.80
6.00
7.00
7.20
7.40
7.60
7.80
Re,(ng/kg)
FIG. 2. (a) Depth profiles of salinity (open circles) and Re (solid circles) measured in this study. (b) Depth profile
of Re normalized to 35460salinity. The vertical line represents the average Re concentration of the samples below 400
m ( 7.42 n&kg), and the shaded area denotes the 95% confidence limits of this value (20.04 ng/ kg).
4102
A. D. Anbar et al.
TABLE 2. Estimates of the average Re concentration in seawater
LDcation
Relno/ka)Reference
Padfic Ocean
7.42 f
0.04
(a) This study
Pacific Ocean
9.1
4.4
(b) Koide et al., 1967
f
Padtic Ocean
6.4
f
2.4
(c) Scadden, 1969
Atlantic Ocean
4.0
f
2.1
(d) Olafsson and Riley, 1972
Atlantic Clcean
6.9
f
2.1
Average Ocean
6.19 k 0.37
(e) Matthews and Riley, 1970
(f) Colodner (1991)
Depth profiles were measured in all studies except (c), which
included only surface samples. In (b) and (e), no data at depths
shallower than 200 m were obtained. Studies (b), (c). (d) and (f)
included samples from more than one location. Uncertainties in (b),
(c) and (1) were reported by the authors as f 1 o of the
measurements, but have been converted to f 20 for comparison
with this study. No uncertainty estimates were included in (d) and
(e); by analogy to (b) and (c), we have calculated their uncertamties
as f 2a. Our average, (a), mcludes all measurements below 400
m, and the uncertainty is reported as f 20 of these values.
Avera 8 values from studies (a) and (f) are reported on a salinitynormakd
basis.
reported on a salinity normalized basis with 2u uncertainty.
There appears to be a systematic offset between our data and
those of COLODNER( 199 1). This discrepancy may in part
be due to the fact that the results of this study are based on
filtered seawater samples, as compared to the unfiltered samples used by COLODNER( 199 1). For example, KRISHNASWAMI et al. ( 198 1) and COCHRANet al. ( 1983, 1990) have
shown that a significant fraction of the *‘(‘Pb in seawater is
bound to particles. This is true for many trace elements (e.g.,
BRULAND,1983; WHITFIELDand TURNER, 1987). However,
if this mechanism alone were to account for the discrepancy
between the data sets, it would indicate a much greater reactivity of Re than is implied by the general uniformity of Re
concentrations with depth. Further work is needed to determine the source of the discrepancy, and whether differences
in acidification or storage procedures are important.
It should be noted that some of the previous studies included data from several locations (see Table 2 ) . Thus, some
of their variability could have been due to real geographical
variations in the Re content of seawater. We believe such
variability is unlikely, however, due to the apparently conservative behavior of Re, as illustrated in Fig. 1, and because,
for the less precise earlier work, there appears to be as much
variability within a depth profile at one location as there is
between samples collected at different sites (see, especially,
KOIDE et al., 1987).
The precision of our data allows detail in the Re profile to
be resolved for the first time. The Re concentration decreases
from a value of 7.34 + 0.04 ng/kg at 4705 m to an average
value of 7.22 + 0.04 at 402 m, and then rises to an average
of 7.69 + 0.07 at 6 m (Fig. 2a). The decrease in the Re
concentration between 4705 and 402 m is comparable to the
percent decrease in salinity over the same depth range
(= l.S%), which is consistent with conservative behavior.
However, at 100 m, the Re concentration does not exceed
the deep-water value, although there is a pronounced salinity
maximum at this depth. A good correlation with salinity
should be reflected in a Re concentration of a7.4 ng/kg at
100 m (Fig. 2b). Although the sampling was not ideally
spaced at these depths, this predicted value is nearly l.O1.5% higher than the average of the measured values at 100
m, and, therefore, should be detectable well within our analytical uncertainty. The salinity-normalized data suggest that
the 5 1 m sample may also be somewhat depleted in Re relative
to the deep waters, although this variation is within the uncertainties of the measurements. The surface enrichment is
reflected only in the 6 m sample, and persists in the salinitynormalized data. The measured Re concentration at the surface is =4% higher than the average of the other samples,
which is well above the analytical uncertainties.
These observations suggest that while the behavior of Re
is essentially conservative, its distribution may be modified
by two types of processes. The first process is a slight depletion
due to biological or inorganic scavenging near the surface,
and subsequent regeneration at depth; this mechanism could
account for the apparent depletion at 100 m. The second
process is an input of possibly anthropogenic Re to the ocean
surface, probably via the atmosphere, which could be reflected
in the surface Re enrichment. This mechanism is analogous
to that observed for lead ( SCHAULEand PATTERSON,198 1) .
The interpretation of small variations in the Re concentration calls for caution, however, due to potential problems
of Re contamination during sample collection, or removal
of dissolved Re from solution during sample storage. Although the possibility of contamination cannot be ruled out,
particularly for samples from shallow depths, the fact that
the samples from 45 and 51 m are identical within errors,
despite being collected using different types of sampling
equipment, indicates that there was little or no Re contamination from the sampling procedure itself. As for Re removal
during storage, it is possible that a small, but detectable,
quantity of Re was lost due to adsorption on the walls of the
polyethylene bottles, or reduction of ReO; following addition
of HCl to these samples (COTTON and WILKINSON, 1988 ) .
This interpretation would demand that a nearly constant
amount of Re was lost from most of the samples, to account
for the uniformity of the salinity-normalized values below
402 m. Although this seems unlikely, it could account for
the apparent enrichment ofthe 6 m sample, which was spiked
aboard ship soon after sample collection, as well as the nonconservative behavior of the 100 m sample. Clearly, further
work is needed to ascertain the importance of such sample
collection and storage processes before these data are interpreted to the limits of the analytical precision.
These problems notwithstanding, our data confirm that
Re is among the more conservative of elements in seawater,
as exemplified by its relatively constant concentration with
depth. These results also demonstrate that the refined chemical techniques described above, coupled to the ID-NTIMS
technique, have the potential to improve substantially our
understanding of the geochemistry of trace elements previously studied only with difficulty. We are currently extending
this approach to the study of several platinum group elements
in seawater.
Acknowledgments-The authors are grateful to D. Karl, R. Lukas,
C. Winn, D. Hebel, and the captain and crew of the R/V Wecoma
for assistance in sample collection and analysis. D. Karl and R. Lukas
Marine geochemistry of rhenium
kindly provided unpublished data from the HOT-33 cruise. We appreciate the helpful comments of K. K. Turekian and a second,
anonymous reviewer. This work has been supported by NASA &fant
NAG 9-43 and NSF arant OCE 9018534. Shin time on the R/V
Wecoma was supported by NSF grants OCE-88&3329 (D. Karl) and
GCE-8717195 (R. Lukas). Division contribution No. 5159 (777).
Editorial handling:G. Fame
REFERENCES
BROOKINSD. G. ( 1986) Rhenium as anal* for fissiogenic technetium: Eh-pH diagram (25°C. 1 bar) constraints. AppZ.Geochem.
1,513-517.
BRULANDK. W. ( 1983) Trace elements in sea-water. In Chemical
Oceanography,Vol. 8 (ed. J. P. RILEY and R. CHESTER), Chap.
45, pp. 157-220. Academic Press.
C~CHI~~N J. K BACON M. P., ~ISHN~WAMI S., and TUREWAN
K. K. I 1983)‘ik’Poand 2’opb~~butio~
in the central and eastern
Indian Gcean. Earth Planet Sci. Lett. f&433-452.
C~CHRAN J. K., MCKIBBIN-VAUGHN T., D~RNBLASER M. M.,
HIR~CHBERGD., LIVINGSTONH. D., and BUE~SELERD. ( 1990)
*“Pb scavenging in the North Atlantic and North Pacific Oceans.
Earth Planet. Sci. Lett. 97, 332-352.
COLODNERD. ( 1991) The marine geochemistry of rhenium, iridium,
and piatinum. Ph.D. thesis, MIT/~OI,
WHOI-91-30.
CWTT~N F. A. and WIL~NSON G. (1988) Advanced Inorganic
Chemistry, 5th ed. J. Wiley & Sons.
CREASERR. A., PAPANASTASSIOU
D. A., and WASSERBURGG. J.
( 199 1) Negative thermal ion mass spectrometry of osmium, rhenium, and iridium. Geochim. Cosmochim. Acta 55,397-401.
DELMOREJ. E. ( 1987) Rare-earth catalyzed oxidation of rhenium
to ReO; and ReO, as observed by ne@ive surface-ionization
rn~-s~omet~.
J. Phys. Chem. 91,2883-2886.
GOLDBERGE. D., KOIDEM., YANG J. S., and BERTINEK. K. ( 1986)
Some comparative marine chemistries of platinum and iridium.
Appi. Geochem. 1,227-232.
GOLDBERGE. D., KOIDEM., YANG J. S., and BERTINEK. K. ( 1988)
Comparative marine chemistries of platinum group metals and
their periodic table neighbors. In Metal Speciation:Theory,Analysis
and A~piicafion (ed. J. R. KRAMERand H. E. AUEN), Chap. 10,
pp. 20 l-2 17. Lewis Publishers.
GRAMLICHJ. W., MURPHY T. J., GARNER E. L., and SHIELD.S
4103
W. R. ( 1973) Absolute isotopic abundance ratio and atomic weight
of a reference sample of rhenium. Nat. Bur. Stds. J. Res. 77A,
691-698.
HEUMANNK. G. ( 1988) Isotope dilution mass spectrometry. In Inorganic Mass Spectrametry (ed. F. ADAMSet al.), Chap. 7, pp.
301-376. J. Wiley & Sons.
HUFFMANE. H., OSWALTR. L., and WILLIAMSL. A. ( 1956) Anionexchange separation of molybdenum and technetium and of tungsten and rhenium. J. Inorg. Nucl. Chem. 3,49-53.
JACINTOG. S. and VAN DEN BERGC. M. G. ( 1989) Dif&ent behavior
of ulatinum in the Indian and Pacific Oceans. Nature 338. 332-
334.
KARL D. M. and WINN C. D. ( 199 1) A sea of change: Monitoring
the oceans’ carbon cycle. Environ. Sci. Technoi. 25, 1976-198l.KOIDE M.. HODGE V.. YANG J. S.. and GOLDBERGE. D. ( 1987)
Determination of rhenium in marine waters and sediments by
graphite furnace atomic absorption spectrometry. Anal. Chem. 59,
1802-1805.
~ISHNASWAM~S., SARINM. M., and ~MAYAJULU B. L. K. ( 198 1)
Chemical and radiochemical investigations of surface and deep
particles of the Indian Ocean. Earth Planet. Sci. Lett. 54,81-96.
LEE D. S. ( 1983) Palladium and nickel in north-east Pacific waters.
Nature 305,47-48.
MATTHEWSA. D. and RILEY J. P. ( 1970) The determination of
rhenium in seawater. Analyt. Chim. Acfa 51,455-462.
OLAFSSON
J. and RILEY J. P. ( 1972) Some data on the marine geochemistry of ~enium. Chem. Geol. 9,227-230.
SCADDENE. M. ( 1969) Rhenium: Its concentration in Pacific Ocean
surface waters. Geochim. Cosmochim. Acta 33,633-637.
SCHAULEB. K. and PATTERSONC. C. ( 198 1) Lead concentrations
in the northeast Pacific: Evidence for global anthropopgenic perturbations. Earth Planet. Sci. Left. 5497-l 16.
V~LKENINGJ.. WALCZYKT.. and HEUMANNK. G. ( 199 1I Osmium
isotope ratio d~~i~tio~s
by negative thermal ion&ion mass
spectrometry. Intl. J. Mass Spectrom. Ion Proc. 105,147- 159.
WA~SERIXJRGG. J., PAPANASTA~SIOU
D. A., NENOW E. V., and
BAUMANC. A. ( 1969) A programmable magnetic field mass spectrometer with on-line data processing. Rev. Sci. Instrum. 40,288295.
WHITFIELDM. and TURNER D. R. ( 1987) The role of particles in
regulating the composition of sea water. In Aquatic Sut$zce Chemistry (ed. W. STUMM), pp. 457-493. J. Wiley & Sons.
ZEININGERH. ( 1984) Ph.D. thesis, University of Regensburg.